8 LangChain Projects to Learn From (Code on GitHub)

Best way to learn LangChain is to read code that actually ships. These 8 repos are the ones we recommend.

1. langchain-ai/langchain (core)

The framework itself. Read the abstractions — Runnable, Chain, Agent. Top-level API has improved dramatically post-0.3.

2. langchain-ai/langgraph

Stateful agents. Study the examples folder — customer_support_bot.py is the best multi-step agent reference.

3. langchain-ai/open-canvas

Real product built with LangGraph. Editing assistant with artifacts. Production patterns for streaming + persistence.

4. assafelovic/gpt-researcher

AI research agent that produces 5K-word reports with citations. Multi-agent orchestration done right.

5. langfuse/langfuse

Observability for LangChain apps. Self-hostable. Read their integration code for how to instrument production apps.

6. langchain-ai/chat-langchain

The chat-with-LangChain-docs assistant. Open-sourced. Reference for production RAG.

7. jerryjliu/llama_index

Not LangChain, but the alternative — worth comparing patterns. LlamaIndex is more opinionated about indexing.

8. langchain-ai/notebooks

Official notebook tutorials. Run them locally — best onboarding path for new LangChain devs.

What to study (in order)

1. Build a basic Runnable chain (prompt → LLM → output parser)
2. Add a RAG retriever
3. Convert to an Agent with one tool
4. Move to LangGraph for stateful flow
5. Add observability (LangSmith or Langfuse)
6. Productionize: error handling, streaming, persistence

Hire a LangChain developer →